The Desmopressin Test Predicts Better Than Basal Cortisol the Long-Term Surgical Outcome of Cushing's Disease.

CONTEXT Cushing's disease (CD) has a significant relapse rate after successful transsphenoidal surgery (TSS). Many CD patients respond aberrantly to the desmopressin test (DT). Disappearance of this response after surgery may suggest complete removal of abnormal corticotrophs and a lower possibility of recurrence. OBJECTIVE The utility of postoperative DT to predict long-term outcome compared to the widely used postoperative cortisol level. DESIGN Retrospective analysis. SETTING Tertiary hospital. PATIENTS Seventy-three patients underwent TSS and postoperative DT; 51 had sustained remission, defined as normal dexamethasone suppression and urinary free cortisol at 6 months. After excluding 12 patients with short follow-up, negative or no preoperative DT, we analyzed 39 patients. INTERVENTION(S) Measurements of morning cortisol at 1-2 weeks and DT within 6 months after TSS. MAIN OUTCOME MEASURE(S) Recurrence or remission at latest follow-up. RESULTS Mean follow-up was 63 ± 50 months. Recurrence occurred in seven patients. In logistic regression analysis, postoperative cortisol levels were not associated with remission. Apart from the percentage increment of cortisol, all other DT criteria (peak cortisol, peak ACTH, absolute cortisol increment [ΔCort], absolute ACTH change, and percentage absolute ACTH change) were significant predictors of outcome. In receiver operating characteristic analysis, the ΔCort had the best diagnostic performance. ΔCort <7.4 μg/dL had a sensitivity of 97% to detect remission. Comparison of Kaplan-Meier curves showed that ΔCort <7.4 μg/dL was associated with remission, whereas ΔCort ≥7.4 μg/dL had a hazard ratio of recurrence of 24.7 (95% confidence interval, 10.6-448.5) at 60 months (median). CONCLUSION Loss of desmopressin response indicates favorable prognosis and, if used in addition to basal cortisol levels, improves the accuracy of the postoperative assessment of CD.